Date of Award
2006
Degree Type
Thesis
Degree Name
Master of Engineering Science
Program
Electrical and Computer Engineering
Supervisor
Dr. Lyndon J. Brown
Abstract
In this thesis, improvements and extensions of an internal model based adaptive algorithm for predictable disturbance identification and cancellation are presented. This algorithm is applied to an active noise cancellation application based on Ben Amara’s acoustic duct model, where the sound signal is composed of a sum of sinusoids. We also extend the algorithm for canceling a disturbance composed of a sum of exponentially damped sinusoidal (EDS) signals. The state variables of an internal model controller in a feedback loop can provide estimates of both parameters, the frequency and the damping factor, of the EDS signals. By using additional integral controllers, the estimation errors can be eliminated. The sinusoidal signal in the active noise cancellation problem, is a special form of the EDS signals with zero damping factor. An internal model is capable of automatically adjusting its frequency to a frequency component of the signal. A new method of generating phase shifted signals is also given. It is also shown how to achieve good noise reduction without the use of an error microphone when a good duct model is available. The convergence of the proposed adaptive algorithm and the stability of the feedback control system including the algorithm are justified using singular perturbation theory and averaging theory. Simulation results demonstrate the validity of the convergence and stability analysis
Recommended Citation
Lu, Jin, "Control and Identification of Narrow-band Disturbances and Signals" (2006). Digitized Theses. 5019.
https://ir.lib.uwo.ca/digitizedtheses/5019
